This application is for a five-year competitive renewal to study the role of the serotonin (5-hydroxytryptamine [5-HT]) system in suicidal behavior in adolescents. Serotonin system dysfunction is related to suicidal behavior in adults at risk for suicide attempts or suicide. Low levels of the serotonin metabolite 5-HIAA in cerebrospinal fluid (CSF) is a biological marker identifying at-risk adults. Therefore, it may also identify suicidal adolescents at risk for future attempts. One hundred additional adolescents, 33 per group, will be recruited to the supplement the 3 groups previously studied. The resulting sample will include 2 groups of inpatient adolescents with major depressive disorder (MDD), matched for age (12-18 years), who will participate in a prospective 24 month follow-up program using standardized ratings to assess depression and suicidal behavior. The study will recruit MDD attempters (N-53) who are consecutively referred for hospitalization because of a suicide attempt during the previous 12 months and age-matched (N-53) MDD non-attempters. These groups will be studied during a 4-8 week psychiatric inpatient hospitalization with a lumbar puncture, platelet 5-HT2 receptor binding, and a fenfluramine challenge test. Age-matched normal adolescents (N-54), with no history (in self or in 1st- and 2nd-degree relatives) of depression and suicidal behavior, will be recruited as controls for fenfluramine challenge and platelet 5-HT2 binding, but not for CSF 5-HIAA. Specific Aim 1 is to determine by prospective follow-up if baseline levels of CSF 5-HIAA, platelet 5-HT2 receptor binding, PRL responses to fenfluramine will predict higher rates of more medically damaging re- attempts at the 3-, 12- and 24-month double blind follow-up assessments. Specific Aim 2 is determine if the enlarged sample of attempters (N-53) will continue to have significantly different prolactin responses to FEN and more 5-HT2 binding sites (BMAX) than depressed non-attempters (N-53) and normal adolescents (N-54); in addition, for the inpatient adolescents only, to determine if the attempter group will have significantly lower baseline levels of CSF 5-HIAA than a group of depressed non-attempters. The attempters with high medical damage will have lower CSF 5-HIAA than those with less medical damage or those in Group 2. Specific Aim 3 is to determine, for the enlarged attempter and non- attempter groups, if the measures of serotonergic function (CSF 5-HIAA concentrations, delta-Peak PRL, 5-HT2 BMAX) will correlate with severity of suicidal behavior, as shown by extent of medical damage among the attempters, and measures of aggression.